Substituting conventional air conditioning systems for cooling with solid desiccant cooling systems (SDCSs) appears to be an interesting alternative for both energy saving and better environment and indoor air quality. Because desiccant wheel (DW) is one of the most important components of SDCSs, the precise prediction of its parameters is vital in the overall performance of the systems. The aim of this investigation is to offer an accurate, robust, and fast modelling approach for the prediction of various parameters of DWs. In this work, a novel hybrid model based on least squares support vector machine (LSSVM) and genetic algorithm (GA) is developed to predict accurately process outlet temperature and humidity (T-pro,T-out and omega(pro,out)), regeneration outlet temperature and humidity (T-reg,T-out and omega(reg,out)), dehumidification effectiveness (eta(deh)), moisture removal capacity (MRC), and sensible energy ratio (SER) for both Silica Gel (WSG) and Molecular Sieve (LT3) materials considering different supply/regeneration section area ratios. The capability of the model was evaluated through three different statistical error tests. The results revealed that integration of LSSVM and GA is a favorable technique for predicting the DWs with a mean average error (MAE) less than 0.23, determination coefficient (R-2) greater than 0.994, and mean squared error (MSE) less than 0.072. (C) 2016 Elsevier Ltd. All rights reserved.